Quantitative structure-toxicity relationship studies of aromatic aldehydes to Tetrahymena pyriformis based on electronic and topological descriptors

Abstract

To establish a quantitative structure-toxicity relationship (QSTR) of a series of 77 aromatic aldehydes for their acute toxicity against Tetrahymena pyriformis, were used the principal component analysis, the multiple linear regression and the multiple nonlinear regression analysis. We proposed linear and nonlinear models and interpreted the toxicity of the compounds by multivariate statistical analysis. The proposed models have been validated using internal validation and external validation techniques and an agreement between experimental and predicted values was verified. The applicability domains of MLR models were investigated using William's plot to detect outliers and outsides compounds. It is interesting to note that some of the selected electronic and topological descriptors in our models are better for the prediction of new similar molecules.